THE VALUE RELEVANCE OF INTANGIBLE ASSETS AND NON-GAAP EARNINGS: EVIDENCE FROM THE BRAZILIAN STOCK MARKET
Intangible Assets. Non-GAAP earnings. Value relevance. EBITDA
Intangible assets are becoming increasingly prominent at the academic and corporate scopes. It is very important for managers and investors to understand their role as part of the organization's value creation system (LEV; DAUM, 2004), as well as sustainability, competitiveness and strategic differential for companies (VILLALONGA, 2004; EHIE; OLIBE, 2010; BHATIA; AGGARWAL, 2016). At the same time, non-GAAP earnings disclosed by press release managers have been gaining prominence over GAAP earnings, due to the continuing loss of relevance of financial information prepared under the Generally Accepted Accounting Principles (LEV, 2018) under the argument that increased non-GAAP earnings informativeness by excluding non-recurring items from earnings (DICHEV; TANG, 2008). In this sense, this research aims to investigate the value relevance of intangible assets and the alternative measure of non-GAAP earnings EBITDA compared to GAAP disclosures of companies listed in [B]³ - Brazil, Bolsa, Balcão, from 2010 to 2018. The EBITDA metric, calculated according to CVM Instruction 527/2012, was chosen because it was the most widely reported non-GAAP measure by Brazilian companies (KPMG, 2016). Survey data will be collected through the Bloomberg® database and standardized financial statements filed on the CVM website. The EBITDA variable will be collected in press releases disclosed by companies. The companies that financial sector were excluded from the study due to differences in the composition of their operations, regulation of financial reporting, sector-specific characteristics (MACHADO; MACEDO; MACHADO, 2015; DUARTE; GIRÃO; PAULO, 2017). To achieve the proposed objective, a comparison will be made between Ordinary Least Squares (OLS), the most used method in value relevance research, and Quantile Regression (QR), supported by the argument that this technique can mitigate some common empirical problems present in estimates in OLS (DUARTE; GIRÃO; PAULO, 2017).